M2 Ultra Chip Benchmark Results Reveal Impressive Performance Gains

This Apple engineer said no DXVK was used: ”We built a dxil to metallib converter and directx11 and directx12 to Metal runtime translator. Non graphics APIs are translated by Wine. But we don’t use any tech from moltenVK or DXVK or spirv-cross etc. The shader converter can be shipped by games and can be used in the game developer asset pipelines.”
Interesting because DXVK is mentioned in their license.
 
The M series chips are impressive, and the Ultra's are overkill for most applications. Where the Ultra's have disappointed was in GPU performance vs a dedicated graphics card. But if Apple's claims hold true, the improved GPU performance on the M2 Ultra should take it into Nvidia territory and finally make it a viable option for 3D graphics work and rendering. Can't wait to see hands-on videos of it with Blender.
Even the stock M2 is a viable option for 3D graphics work and rendering. I am in this industry and what people are forgetting is the actual relevant work-flow. If you only reach the 'cap' of your machine's capabilities a few minutes a day (e.g the final render, or a viewport with a lot of complex shading/geometry), most professionals don't bother, especially when the actual final output is rendered externally. A lot of the time, you are working on a detail, and the environment is mostly 'idle'.
Ironically, mostly gamers will complain, because they demand constant peak performance on the GPU side.
 
Mac Pro 2023 is totally messed up.
Depends on your perspective. I see it as a placeholder, required because the MP needed updating and perhaps reaffirmation that the MP has a path forward for those interested.

IMHO, it's likely that the original plan was more ambitious, and had to be replaced when the necessary chips did not materialize in time.

I expect a fairly rapid update as soon as the chips arrive, maybe early next year.
 
Because they are way more powerful and faster. Why do you ignore that?
Brah… listen to yourself. Let me give you analogy so you can see how silly this argument is.

“My grandmas 1986 Wagon is faster than your dad’s 2023 Ram.”

“How so? my dad’s Ram outperforms that old wagon, it was made to tow trailers, it has great torque, should blow granny off the line”

“Um… no way my grandma is awesome! I saw a show on TV where some guy that owned his own garage, redid the engine in an old wagon, and now it can beat Mustangs!”

It’s ridiculous childs talk. We are talking about real items for sale. Not a hobby project. You can’t have an honest discussion if you bring in “what abouts” and sillyness.
 
The choice to ditch Windows for macOS. 😁
Did you notice how desperate Mac owners are to be able to run PC software: Linux, games, MS office you name it. Hence all sorts of emulation etc. You never see PC owners trying to figure out how to run Mac software. Does not it tell us how inferior MacOS ecosystem is? Why would anyone want to switch to MacOS and then dream of running the software that is available on PCs?
 
So we are comparing an M2 Ultra with Intel 28 Core unit from 2016, which seems unfair.

Except it is 2019.

Released Q2 '19 at list price of $8198 with a "> 1TB RAM" tax on it.


while there was another 28 core option that was $4894 ( is $3,304 less expensive)

Chop that RAM tax off and the 28 core MP model could have been $7,995. Apple is bringing a huge price drop by just dropping the sky high tax they threw at the Intel model. ( Apple has their 20% markup on top of the Intel tax. )
Those upper two models of the MP 2019 have been 'alternative universe' priced for years. Even Intel gave up on that excessive RAM tax in 2021 for the W-3300 series. (that almost no one bought).

Definitely, one place where AMD brought some competitive sanity to the market. ( no extra mark up on "Pro" TR models for crossing the TB threshold. ). Intel's RAM tax was a greedy money grab because they knew at that point that 10nm was going off the tracks and they'd need tons of money to 'fix' the problem. AMD under cut them on price and over time delivered better performance ( and took a huge chunk of market share away). Intel's plan was short term greedy ( would look good for 2-3 financial quarters , and long term dumb as a rock ( huge revenue losses in the long term). )


Apple put better 'value' into the $6999 price point , but they kind of set that Intel 28 core model up for 'failure' on purpose. Probably so they could be the 'hero' now.


These days Intel and AMD have CPUs with 56 and 64 cores with access to 4TB of RAM.

Not particularly material for macOS since it doesn't do more than 64 threads. With SMT those 56 and 64 cores are 112 and 128 threads. Way , way , way past what macOS can cover. Apple wouldn't have picked those anyway. Pragmatically really talking about 32 core limits for viable x86_64 options unless going to set the firmware to turn off SMT completely.

If Apple was going to use an Intel model now it likely would be a W-2400 that has about the similar limit as before ( 24 cores )

The major reason that Apple really wants app developers to adjust their software to M-series specific optimizations is Apple isn't looking to crank the CPU thread count over 64. They are looking to push more 'embarrassingly parallel' work of to AMX , NPU cores , GPU cores , etc that don't have the CPU thread limit. It is probably a good idea for a single user workstation. Those server package 56-64-96 core counts are really mostly targeting multiple user / multiple tenant where there are lots of concurrent folks doing different things on the same hardware.


The most significant benefit of the M processors is power consumption, but workstation has fewer restrictions and needs to deliver raw power.

Not really. Apple's single thread is better than many of these extremely high core count x86 packages because there aren't substantive wasted thermal bleed from the rest of the chip.



I would love to see a comparison with the newer CPUs from AMD and Intel.

In single threaded the M2 Ultra beats w9-3495X and Threadripper W5595X but looses to Core i9-13900K


[ The Threadripper W5595X is getting somewhat old. About 2 years old. The W7000 series should be coming by end of the year and Apple should be nervous about that one. ] . That "power savings" is a win for a single user workstation with a mix of single/multiple threaded apps running.


In multiple threaded stuff it looses to the server "hand me down" chips , but very close to the Core i9-13900K (win some / loose some ). So it is something kind of in the middle between a 'top fuel dragster' consumer chip and a server multiple user chip. For a single user workstation it is probably the right call. It isn't 'bad' at all. It isn't a Xeon or Threadripper 'killer' solution , but it really doesn't have to be.
 
20% is not enough of a reason for me (as a casual user) to upgrade my M1 Mac Studio Ultra to an M2 Mac Studio Ultra but I can see why businesses may well consider if if they can complete tasks 20% faster. Time is money and it won't take long to recuperate the cost of upgrading.
If you are a casual user, why did you even choose the M1 Ultra in the first place?
 
Looking at that Geekbench page, I notice that there are stats from 3.67GHz and 3.0GHz machines. That’s a pretty big clock difference. Are they running 3.67 in the Pro and 3.0 in the Studio, or is the higher-spec chip with the bigger GPU also running at a higher clock?
 
Based on power consumption Apple M series chips outperform everything Intel has to offer. That's also reality, right? Just not the one that's convenient for you.
Intel has many chip series optimized for different things: power consumption, performance and everything in between. Apple only has one chip type (with different core counts).
 
I still don’t get why Apple did not differentiate the Mac Pro anymore from the Mac Studio. Using the identical CPU there isn’t much reason to spend the much more money. Sure you get PCIe expansion, but you can get that in a TB4 Sonnet expansion box at a quarter the cost.

And which point you can stop on buying SSD capacity from Apple at the 1TB ( or maybe 2TB ) level for os/app/home directory and buy much more market rate priced SSD capacity. If need 8-16TB of internal drive space the Mac Pro is cheaper. It is what you do with the slots.


You can’t use the PCIe for graphics cards, and you are still limited on the RAM…

There are over 50+ PCI-e cards that work in the Mac Pro. That is what Apple is doing a terrible sales job of right now. Most folks are mapping GPU card == PCI-e card in their head and saying 'nothing' is really different here. That is just shallow, but it shouldn't be surprisingly shallow for Apple's marketing/sales team to help customers find their way out of with some information. Lots of things point to Apple being in the same 'pit' in their outlook.
( I suspect though they also don't want to drag too much attention away from driving folks to buying 2-8TB of SSD capacity directly from them at highly inflated prices. moooore money, mo money , mo money ) .

If someone needs high capacity and/or high write overhead working set 'scratch' drive then Mac Pro is way better than a Studio. Better than 10GbE ( if primary network interface is WiFi 5 then yeah not much difference). Same for any x8 or x16 cards (there are lots of those outside of GPUs).
 
Last edited:
The choice to ditch Windows for macOS. 😁
How do you play MFS (Microsoft Flight Simulator) on macOS? And what about people who depends on CUDA and Nvidia acceleration (like on Matlab)? Software compatibility is first thing on priority list.
 
Did you notice how desperate Mac owners are to be able to run PC software: Linux, games, MS office you name it. Hence all sorts of emulation etc. You never see PC owners trying to figure out how to run Mac software. Does not it tell us how inferior MacOS ecosystem is? Why would anyone want to switch to MacOS and then dream of running the software that is available on PCs?
The only people who buy Office for Mac are former Windows users, who feel they can't let go. They also long for window snapping tools and antivirus software. macOS is a proper Unix, so I don't know why you would want to run Linux? I tell you what I dream of, I want to replace JDownloader for a native macOS app. Cross-platform apps are the most horrible thing I can think of. Nobody wants to run them, sometimes you just can't avoid.
 
If you had three 24 inch monitors and you were contemplating getting a desktop Mac, which would you choose or recommend: base model Mac Mini M2 or base model Mac Studio M2/M1? I feel like a Mac Mini base model or M2 Pro would struggle driving 3 monitors in addition to daily usage.
 
How do you play MFS (Microsoft Flight Simulator) on macOS? And what about people who depends on CUDA and Nvidia acceleration (like on Matlab)? Software compatibility is first thing on priority list.
Not for Mac users. Software quality over software compatibility!
 
I have M1 based computers that are performing better than I had expected so I don't feel any need to upgrade.

But, I like the direction all this seems to be headed.
 
Did you notice how desperate Mac owners are to be able to run PC software: Linux, games, MS office you name it. Hence all sorts of emulation etc. You never see PC owners trying to figure out how to run Mac software. Does not it tell us how inferior MacOS ecosystem is? Why would anyone want to switch to MacOS and then dream of running the software that is available on PCs?
That's a dumb analysis. A better analysis is

(a) one of the drivers (not the only, but one of the drivers) for Vision Pro uptake will be games. Hence there's value for Apple to get some top tier games on the platform for when it is released. They probably won't have to try to hard to persuade game companies -- after all the developers themselves will want their games on this hardware -- but the cheaper and faster you can make the porting process, the better.

(b) Apple is at the point where they clearly have a superior product across the entire PC line. So why do they not have market share at 50%? 70%? 90%? It's time to start tackling the issues that prevent various niches from moving over. They won't compromise the product to reach everyone, but they might as well make minor changes if that allows an additional 10% market share to move over. Games is an example of such a situation. Some things (like external GPUs) will not be supported because they compromise the long term plan; but other things, like just making it easier to port games, why not?
 
Web developers would love to run Safari for testing, and app developers would love to build iOS apps.

Beyond that, there have absolutely been Mac apps some Windows users are jealous of.
You are talking about cross platform development. That's a different story (and BTW, web developers care about Safari very little because Safari market share is really low). Sure there are some nice to have apps on Macs but there are equally good apps in the same domain on PCs. Whereas Macs simply can't run some must have apps. Hence constant discussions about how to emulate/port Windows apps on Macs (including lots of discussions here on MR) and nothing of sorts among PC users.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Back
Top